Özet:

We propose error correction iteratively decodable cyclic codes (IDCC) that consist of two cyclic Hamming codes with different generator polynomials. As a mathematical apparatus, we apply the theory of linear finite-state machines (LFSM) in binary Galois fields. A generalized decoding algorithm was constructed based on power permutation of bits in the code word and the new technique for combining the codes. By using hard decisions only, it is possible to achieve high speed and simple hardware-software implementation of encoder and decoder on linear feedback shift registers. The IDCC (n, k)-code makes it possible to correct the errors of multiplicity to (n−k). A code word may have arbitrary length: both small and large. Code rate (k/n) is close to one. It was established in the course of research that approaching the theoretical limit (border) by Shannon maximally close significantly increases length of codes, complicates encoders and decoders, increases a delay in decoding, and other problems appear. That is why the main criterion for the optimality of error correction coding is proposed to be those code characteristics that are important for practical application (time and hardware costs, focus on contemporary circuitry and parallel processing). From this point of view, the developed IDCC codes can be considered as an alternative to well-known iterative codes (LDPC codes and turbo codes) whose main advantage is the maximum proximity to the Shannon limit. This is important because at the present stage of development of science and technology one of the relevant scientific and technological problems is the task on ensuring high reliability of data transmission in different systems of digital communication. The proposed codes make it possible to solve the specified task at minimal resource costs and high efficiency. Author Biography Vasyl Semerenko, Vinnytsia National Technical University Khmelnytske highway, 95, Vinnytsia, Ukraine, 21021 PhD, Associate Professor Department of computer technique

Anahtar Kelimeler: